Research Experiences in Parallel Numerical Linear Algebra

并行数值线性代数研究经历

• 批准号: 9987895
• 负责人: Daniel Okunbor
• 金额: $ 3万
• 依托单位: University of Maryland Eastern Shore
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2000
• 资助国家: 美国
• 起止时间: 2000-09-01 至 2001-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9987895&HistoricalAwards=false
• 关键词: Research; Experiences; Parallel; Numerical; Linear

The goal of this project is to take selected undergraduate students, who have had limited or no opportunities for research, and provide them with a significant research experience that will stimulate their interest in mathematical and computational sciences to the point that they will react positively to the possibility of attending graduate school for further education and research in mathematics realted areas. The project is selected to illustrate methods of research that will take the students progressively from the definition of the problem, though the statement of objectives, literature search, and the actual research, to the point of writing up their results and conclusions.In this project, students will be given an opportunity to do research in the area of numerical algorithms for problems arising from different areas in science and engineering. In particular, students will be involved in the design, analysis, and implementation of parallel numerical algorithms for the solution of large sparse linear systems, eigenvalue, least squares, optimization problems. This research will involve investigating the current literature in these areas as well as the parallel architecturs that can be utilized for this project. In the instructional phase of the project, students will also learn mathematical concepts in numerical linear algebra, message-passing interface (MPI) and C/C++ programming language with appropriate language extensions to provide a means of implementing the parallel numerical algorithms on parallel processors.

该项目的目标是采取选定的本科生，谁有有限的或没有研究的机会，并为他们提供一个显着的研究经验，这将刺激他们在数学和计算科学的兴趣点，他们将积极响应参加研究生院的可能性，在数学相关领域的进一步教育和研究。 该项目的选择是为了说明研究方法，将学生从问题的定义，通过目标的声明，文献检索，和实际研究，逐步到写他们的结果和结论的点。在这个项目中，学生将有机会做研究领域的数值算法产生的问题，在科学和工程的不同领域。 特别是，学生将参与大型稀疏线性系统，特征值，最小二乘，优化问题的解决方案的并行数值算法的设计，分析和实施。 这项研究将涉及调查这些领域的现有文献以及可用于该项目的并行架构。 在项目的教学阶段，学生还将学习数值线性代数，消息传递接口（MPI）和C/C++编程语言中的数学概念，并提供适当的语言扩展，以提供在并行处理器上实现并行数值算法的方法。